How to Get documents from Index or Alias — ElasticSearch Connector for Azure Data Factory (Pipeline)

Introduction

In this article we will delve deeper into ElasticSearch and Azure Data Factory (Pipeline) integration, and will learn how to get documents from index or alias. We are continuing from where we left off. By this time, you must have installed ODBC PowerPack, created ODBC Data Source, and configured authentication settings in your ElasticSearch account .

So, let's not waste time and begin.

Use Query Builder to generate SQL query

The first thing you have to do is open Query Builder:

ZappySys API Driver - ElasticSearch

Read / write ElasticSearch data inside your app; perform many ElasticSearch operations without coding, just use easy to use high performance API Connector for ElasticSearch

ElasticsearchDSN
Then simply select the Get documents from Index or Alias endpoint (action).
Continue by configuring the Required parameters. You can also set optional parameters too.
Move on by hitting Preview Data button to preview the results.

If you see the results you need, simply copy the generated query:

Get documents from Index or Alias

Required Parameters
Index	Select the value from the dropdown
Optional Parameters
Enter Query (JSON Format)	{"match_all": { } }
CursorTimeout (Minutes)	10
How many Record to fetch per request	1000
Alias (Deprecated - Use Index instead)
Advanced Properties
Filter	$.hits.hits[*]
NextUrlAttributeOrExpr	$._scroll_id
StopIndicatorAttributeOrExpr	$.hits.hits[0]._id
NextUrlEndIndicator	regex=^$
EnablePageTokenForBody	True
HasDifferentNextPageInfo	True
PagePlaceholders
FirstPageBodyPart
NextPageBodyPart
MetaDetectionOrder	Default
DoNotIndentArray	True
Ignore certificate related errors

SELECT * FROM MyIndexOrAliasName --WITH(Query='{"match": { "PartNumber" : "P50" } }')

That's it! You can use this query in Azure Data Factory (Pipeline).

Let's not stop here and explore SQL query examples, including how to use them in Stored Procedures and Views (virtual tables) in the next steps.

SQL query examples

Use these SQL queries in your Azure Data Factory (Pipeline) data source:

How to Read ElasticSearch documents from Index (all or with filter)

Gets documents by index name (i.e. Table name) or alias name (i.e. View name). Using WHERE clause invokes client side engine so try to avoid WHERE clause and use WITH clause QUERY attribute. Use search endpoint instead to invoke query.

SELECT * FROM MyIndexOrAliasName --WITH(Query='{"match": { "PartNumber" : "P50" } }')

How to Read ElasticSearch documents from Alias (all or with filter)

SELECT * FROM MyIndexOrAliasName --WITH(Query='{"match": { "PartNumber" : "P50" } }')

How to Search documents from Index using ElasticSearch Query language

Below example shows how to search on a comment field for TV word anywhere in the text for Index named MyIndexOrAliasName (it can be index name or alias name). For more information on ElasticSearch Query expression check this link https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-match-query.html

SELECT * FROM MyIndexOrAliasName WITH(Query='{"match": { "comment" : "TV" } }')
		
		--or use below - slight faster (avoids table / alias list validation)
		--SELECT * FROM search WITH(Index='MyIndexName', Query='{"match": { "comment" : "TV" } }')
		--SELECT * FROM search WITH(Index='MyIndexName', Alias='MyAliasName', Query='{"match": { "comment" : "TV" } }')

How to Search documents from Alias using ElasticSearch Query language

Below example shows how to search on Alias rather than Index name. Alias is build on index (consider like a view in RDBMS). This example filtes data from Alias with some condition in the Query Text. For more information on ElasticSearch Query expression check this link https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-match-query.html

SELECT * FROM MyAliasName WITH(Query='{"match": { "comment" : "TV" } }')
		
		--or use search endpoint then you must supply both Index name and Alias name 
		--calling /search endpoint in FROM clause is slight faster (avoids table / alias list validation)
		--SELECT * FROM search WITH(Index='MyIndexName',Index='MyAliasName', Query='{"match": { "comment" : "TV" } }')

How to Read ElasticSearch documents from Index (all or with filter)

SELECT * FROM MyIndexOrAliasName --WITH(Query='{"match": { "PartNumber" : "P50" } }')

How to Read ElasticSearch documents from Alias (all or with filter)

SELECT * FROM MyIndexOrAliasName --WITH(Query='{"match": { "PartNumber" : "P50" } }')

How to Search documents from Index using ElasticSearch Query language

SELECT * FROM MyIndexOrAliasName WITH(Query='{"match": { "comment" : "TV" } }')
		
		--or use below - slight faster (avoids table / alias list validation)
		--SELECT * FROM search WITH(Index='MyIndexName', Query='{"match": { "comment" : "TV" } }')
		--SELECT * FROM search WITH(Index='MyIndexName', Alias='MyAliasName', Query='{"match": { "comment" : "TV" } }')

How to Search documents from Alias using ElasticSearch Query language

SELECT * FROM MyAliasName WITH(Query='{"match": { "comment" : "TV" } }')
		
		--or use search endpoint then you must supply both Index name and Alias name 
		--calling /search endpoint in FROM clause is slight faster (avoids table / alias list validation)
		--SELECT * FROM search WITH(Index='MyIndexName',Index='MyAliasName', Query='{"match": { "comment" : "TV" } }')

get_documents endpoint belongs to [Dynamic Table] , [Dynamic Table] table(s), and can therefore be used via those table(s).

Stored Procedures and Views

Create Custom Stored Procedure

You can create procedures to encapsulate custom logic and then only pass handful parameters rather than long SQL to execute your API call.

Steps to create Custom Stored Procedure in ZappySys Driver. You can insert Placeholders anywhere inside Procedure Body. Read more about placeholders here

Go to Custom Objects Tab and Click on Add button and Select Add Procedure:
Enter the desired Procedure name and click on OK:
Select the created Stored Procedure and write the your desired stored procedure and Save it and it will create the custom stored procedure in the ZappySys Driver:
Here is an example stored procedure for ZappySys Driver. You can insert Placeholders anywhere inside Procedure Body. Read more about placeholders here
```
CREATE PROCEDURE [usp_get_orders]
    @fromdate = '<<yyyy-MM-dd,FUN_TODAY>>'
 AS
    SELECT * FROM Orders where OrderDate >= '<@fromdate>';
```
That's it now go to Preview Tab and Execute your Stored Procedure using Exec Command. In this example it will extract the orders from the date 1996-01-01:
```
Exec usp_get_orders '1996-01-01';
```
Let's generate the SQL Server Query Code to make the API call using stored procedure. Go to Code Generator Tab, select language as SQL Server and click on Generate button the generate the code.
As we already created the linked server for this Data Source, in that you just need to copy the Select Query and need to use the linked server name which we have apply on the place of [MY_API_SERVICE] placeholder.
```
SELECT * FROM OPENQUERY([LS_TO_ELASTICSEARCH_IN_GATEWAY], 'EXEC usp_get_orders @fromdate=''1996-07-30''')
```
Now go to SQL served and execute that query and it will make the API call using stored procedure and provide you the response.

Create Custom Virtual Table

ZappySys API Drivers support flexible Query language so you can override Default Properties you configured on Data Source such as URL, Body. This way you don't have to create multiple Data Sources if you like to read data from multiple EndPoints. However not every application support supplying custom SQL to driver so you can only select Table from list returned from driver.

If you're dealing with Microsoft Access and need to import data from an SQL query, it's important to note that Access doesn't allow direct import of SQL queries. Instead, you can create custom objects (Virtual Tables) to handle the import process.

Many applications like MS Access, Informatica Designer wont give you option to specify custom SQL when you import Objects. In such case Virtual Table is very useful. You can create many Virtual Tables on the same Data Source (e.g. If you have 50 URLs with slight variations you can create virtual tables with just URL as Parameter setting.

Go to Custom Objects Tab and Click on Add button and Select Add Table:
Enter the desired Table name and click on OK:
And it will open the New Query Window Click on Cancel to close that window and go to Custom Objects Tab.
Select the created table, Select Text Type AS SQL and write the your desired SQL Query and Save it and it will create the custom table in the ZappySys Driver:
Here is an example SQL query for ZappySys Driver. You can insert Placeholders also. Read more about placeholders here
```
SELECT
  "ShipCountry",
  "OrderID",
  "CustomerID",
  "EmployeeID",
  "OrderDate",
  "RequiredDate",
  "ShippedDate",
  "ShipVia",
  "Freight",
  "ShipName",
  "ShipAddress",
  "ShipCity",
  "ShipRegion",
  "ShipPostalCode"
FROM "Orders"
Where "ShipCountry"='USA'
```
That's it now go to Preview Tab and Execute your custom virtual table query. In this example it will extract the orders for the USA Shipping Country only:
```
SELECT * FROM "vt__usa_orders_only"
```
Let's generate the SQL Server Query Code to make the API call using stored procedure. Go to Code Generator Tab, select language as SQL Server and click on Generate button the generate the code.
As we already created the linked server for this Data Source, in that you just need to copy the Select Query and need to use the linked server name which we have apply on the place of [MY_API_SERVICE] placeholder.
```
SELECT * FROM OPENQUERY([LS_TO_ELASTICSEARCH_IN_GATEWAY], 'EXEC [usp_get_orders] ''1996-01-01''')
```
Now go to SQL served and execute that query and it will make the API call using stored procedure and provide you the response.

Get documents from Index or Alias in Azure Data Factory (Pipeline)

Go to Linked services section and create a new Linked service based on ODBC:
Select "ODBC" service:
Configure new ODBC service. Use the same DSN name we used in the previous step and copy it to Connection string box:

DSN=
For created ODBC service create ODBC-based dataset:
Go to your pipeline and add Copy data connector into the flow. In Source section use OdbcDataset we created as a source dataset:
Then go to Sink section and select a destination/sink dataset. In this example we use precreated AzureBlobStorageDataset which saves data into an Azure Blob:
Finally, run the pipeline and see data being transferred from OdbcDataset to your destination dataset: